From: SourceForge.net <no...@so...> - 2004-02-27 20:33:04
|
Bugs item #906143, was opened at 2004-02-27 15:30 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=112694&aid=906143&group_id=12694 Category: agent Group: None Status: Open Resolution: None Priority: 5 Submitted By: Robert Story (rstory) Assigned to: Stefan Radman (sradman) Summary: subagent get-next/scope violations Initial Comment: Not quite sure if this is a bug, so this is a reminder to self to check it out. Observed on 5.0.8, so may have already been fixed. Say an subagent registers for 1..n instances. When a particular request arrives, 1..(n-1) are unavailable, and the handler will return SNMP_NO_ERROR w/out setting an error for the varbind. What looks wrong to me, is that the master agent will send the request to the subagent, which will hit each instance, 1-(n-1), in turn, until it n returns a values. The subagent then spits out a scope violation error, and returns to the master. the subagent then gets a request starting at 2, and will search 2-n, hitting a scope violation at the end. Then it moves on to 3, etc etc., until it gets to n, which successfully returns a value for n. so, the question is, why doesn't the subagent check scope during the subtree processing to avoid sending requests to out of scope handlers? It would still take n messages from agent to subagent, but each message would only process it's own instance, and not ((n-1) - start). ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=112694&aid=906143&group_id=12694 |